Users of relational databases have existing tables of data that contain obsolete columns. The dropping of a column from a table encompasses two main parts. The first part is the modification of the metadata, i.e., catalog definition, associated with the table's schema definition, and the second part is the reorganization of the table's data to remove the dropped column data. In existing approaches, the table is placed in a restrictive state for the reorganization, which means that the data is unavailable between the time the metadata is modified to reflect the dropped column and the time that the table data reorganization is completed. For tables with substantial amounts of data, the process of reorganizing may result in an unacceptable outage.